A sparsity-based approach for spectral image target detection from compressive measurements acquired by the CASSI architecture.

Hyperspectral imaging requires handling a large amount of multidimensional spectral information. Hyperspectral image acquisition, processing and storage are computationally and economically expensive and in most cases, slow processes. In recent years, optical architectures have been developed for acquisition of spectral information in compressed form by using a small set of measurements coded by a spatial modulator. This work formulates a processing scheme that allows the measurements acquired by such compressive sampling systems to be used to perform spectral detection of targets by adapting traditional detection algorithms for use in the compressive sampling model and shows that the performance is comparable with that obtained by detection processes without compression. [ABSTRACT FROM AUTHOR]